Semi-Supervised Learning via Compact Latent Space Clustering.
File(s)
Accepted version
Author(s)
Type
Conference Paper
Abstract
We present a novel loss function for semi- supervised learning of neural networks with a simple and effective regularization term based on compact clustering of the latent feature space. The key idea is to dynamically create a graph over both labeled and unlabeled training samples using Label Propagation (LP) to capture the underlying structure in the feature space and model its high and low density areas. The regularization attracts similar samples to form compact clusters and repulses dissimilar ones without applying strong forces to unconfident samples. Label confidence is directly obtained via LP in contrast to using predictions from an imperfect classifier as in previous work. We evaluate our approach on three benchmarks and compare to state-of-the art with promising results. Our method can be easily applied to any existing network architecture enabling an effective use of unlabeled data for a wide range of applications.
Date Issued
2018-07-10
Date Acceptance
2018-05-11
Citation
CoRR, 2018, abs/1806.02679, pp.2464-2473
Publisher
PMLR
Start Page
2464
End Page
2473
Journal / Book Title
CoRR
Volume
abs/1806.02679
Copyright Statement
© 2018 The Author(s)
Sponsor
Engineering & Physical Science Research Council (E
Commission of the European Communities
Identifier
https://arxiv.org/abs/1806.02679
Grant Number
RTJ13261760-1
H2020 - 757173
Source
International Conference on Machine Learning
Subjects
cs.LG
cs.CV
cs.NE
stat.ML
Publication Status
Published
Start Date
2018-07-10
Finish Date
2018-07-15
Coverage Spatial
Stockholmsmässan, Stockholm, Sweden